Z80 Refugee wrote: ↑
Sat Jun 02, 2018 8:31 am
Yes, I'm a retired electronics design engineer with military, automotive, and high-reliability systems experience.
My career was somewhat different - TV outside broadcast.
Unlike studio engineers, we had to solve problems on the fly, often without ready-made solutions available on the market.
And it had to work reliably first time - we couldn't ask the horses to re-run the Epsom Derby.
On the other hand, it probably wasn't going to be useful more than once, and wasn't going into production, because by next year's Derby a whole new generation of equipment would be available on the market, and the production team would want to stretch the possibilities in new directions.
I can well understand your attitude - military and automotive systems demand more than a one-off solution, and medical systems take it one stage further.
So before recommending my suggestion, you would have to do a lot more research, and your time would be better spent on improving the power supply.
On the other hand, for a one-off, my idea still stands - though to make it work you'd have to work out the values of the cap and resistor (requiring measurement of what the regulator can supply, how long the dip in voltage lasts, how long between dips, etc.) and you'd have to look for an appropriate diode. Maybe there isn't one - I haven't looked, I'm leaving that to whoever's considering it - but if, as you say, there isn't one that cuts the mustard, we could consider an 'ideal diode' such as appeared in the Pi B+. That would solve the three issues you high-lighted - switch-on time, voltage drop, and current-carrying capacity, though at the expense of greater complexity.